ontology

Build a typed knowledge graph for agent memory by creating and linking entities with defined properties. This enables sophisticated reasoning and data sharing by representing complex information as connected nodes with enforced schema constraints.

Entity management in a typed knowledge graph works via CLI scripts that perform CRUD operations on typed objects. You create entities with defined properties, enforce schema validation, and manage structured data records for complex projects.

Query relationships in a knowledge graph by using CLI scripts to map and retrieve connected objects. You can define specific relations between entities and query for linked items, such as finding all tasks assigned to a person and blocked by another task.

You can enforce schema validation on knowledge graph data by defining types, properties, and relations. This constraint enforcement ensures data integrity by validating all entities and relationships against the defined schema during creation and updates.

You do not need a separate graph database to use structured memory for composable skills. This system operates via standalone CLI scripts for CRUD and querying, managing the typed knowledge graph and multi-step action planning directly through schema validation.

Plan multi-step actions as graph transformations by representing each step as a modification to your typed knowledge graph. Use the CLI scripts to update entity states and relationships, ensuring the action sequence maintains schema constraints and data integrity.